Journal articles on the topic 'Gittens index'
To see the other types of publications on this topic, follow the link: Gittens index.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 journal articles for your research on the topic 'Gittens index.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse journal articles on a wide variety of disciplines and organise your bibliography correctly.
1
Wu, Xianyi, and Xian Zhou. "Open Bandit Processes with Uncountable States and Time-Backward Effects." Journal of Applied Probability 50, no. 2 (June 2013): 388–402. http://dx.doi.org/10.1239/jap/1371648948.
Full textAbstract:
Bandit processes and the Gittins index have provided powerful and elegant theory and tools for the optimization of allocating limited resources to competitive demands. In this paper we extend the Gittins theory to more general branching bandit processes, also referred to as open bandit processes, that allow uncountable states and backward times. We establish the optimality of the Gittins index policy with uncountably many states, which is useful in such problems as dynamic scheduling with continuous random processing times. We also allow negative time durations for discounting a reward to account for the present value of the reward that was received before the present time, which we refer to as time-backward effects. This could model the situation of offering bonus rewards for completing jobs above expectation. Moreover, we discover that a common belief on the optimality of the Gittins index in the generalized bandit problem is not always true without additional conditions, and provide a counterexample. We further apply our theory of open bandit processes with time-backward effects to prove the optimality of the Gittins index in the generalized bandit problem under a sufficient condition.
2
Wu, Xianyi, and Xian Zhou. "Open Bandit Processes with Uncountable States and Time-Backward Effects." Journal of Applied Probability 50, no. 02 (June 2013): 388–402. http://dx.doi.org/10.1017/s0021900200013437.
Full textAbstract:
Bandit processes and the Gittins index have provided powerful and elegant theory and tools for the optimization of allocating limited resources to competitive demands. In this paper we extend the Gittins theory to more general branching bandit processes, also referred to as open bandit processes, that allow uncountable states and backward times. We establish the optimality of the Gittins index policy with uncountably many states, which is useful in such problems as dynamic scheduling with continuous random processing times. We also allow negative time durations for discounting a reward to account for the present value of the reward that was received before the present time, which we refer to as time-backward effects. This could model the situation of offering bonus rewards for completing jobs above expectation. Moreover, we discover that a common belief on the optimality of the Gittins index in the generalized bandit problem is not always true without additional conditions, and provide a counterexample. We further apply our theory of open bandit processes with time-backward effects to prove the optimality of the Gittins index in the generalized bandit problem under a sufficient condition.
3
Banks, Jeffrey S., and Rangarajan K. Sundaram. "Switching Costs and the Gittins Index." Econometrica 62, no. 3 (May 1994): 687. http://dx.doi.org/10.2307/2951664.
Full text
4
Glazebrook, K. D., and R. W. Owen. "Gittins-index heuristics for research planning." Naval Research Logistics 42, no. 7 (October 1995): 1041–62. http://dx.doi.org/10.1002/1520-6750(199510)42:7<1041::aid-nav3220420705>3.0.co;2-e.
Full text
5
Crosbie, J. H., and K. D. Glazebrook. "Evaluating policies for generalized bandits via a notion of duality." Journal of Applied Probability 37, no. 2 (June 2000): 540–46. http://dx.doi.org/10.1239/jap/1014842557.
Full textAbstract:
Nash's generalization of Gittins’ classic index result to so-called generalized bandit problems (GBPs) in which returns are dependent on the states of all arms (not only the one which is pulled) has proved important for applications. The index theory for special cases of this model in which all indices are positive is straightforward. However, this is not a natural restriction in practice. An earlier proposal for the general case did not yield satisfactory index-based suboptimality bounds for policies — a central feature of classical Gittins index theory. We develop such bounds via a notion of duality for GBPs which is of independent interest. The index which emerges naturally from this analysis is the reciprocal of the one proposed by Nash.
6
Crosbie, J. H., and K. D. Glazebrook. "Evaluating policies for generalized bandits via a notion of duality." Journal of Applied Probability 37, no. 02 (June 2000): 540–46. http://dx.doi.org/10.1017/s0021900200015722.
Full textAbstract:
Nash's generalization of Gittins’ classic index result to so-called generalized bandit problems (GBPs) in which returns are dependent on the states of all arms (not only the one which is pulled) has proved important for applications. The index theory for special cases of this model in which all indices are positive is straightforward. However, this is not a natural restriction in practice. An earlier proposal for the general case did not yield satisfactory index-based suboptimality bounds for policies — a central feature of classical Gittins index theory. We develop such bounds via a notion of duality for GBPs which is of independent interest. The index which emerges naturally from this analysis is the reciprocal of the one proposed by Nash.
7
El Karoui, N., and I. Karatzas. "General Gittins index processes in discrete time." Proceedings of the National Academy of Sciences 90, no. 4 (February 15, 1993): 1232–36. http://dx.doi.org/10.1073/pnas.90.4.1232.
Full text
8
Bank, Peter, and Christian Küchler. "On Gittins’ index theorem in continuous time." Stochastic Processes and their Applications 117, no. 9 (September 2007): 1357–71. http://dx.doi.org/10.1016/j.spa.2007.01.006.
Full text
9
Weber, Richard. "On the Gittins Index for Multiarmed Bandits." Annals of Applied Probability 2, no. 4 (November 1992): 1024–33. http://dx.doi.org/10.1214/aoap/1177005588.
Full text
10
Glazebrook, K. D., D. Ruiz-Hernandez, and C. Kirkbride. "Some indexable families of restless bandit problems." Advances in Applied Probability 38, no. 3 (September 2006): 643–72. http://dx.doi.org/10.1239/aap/1158684996.
Full textAbstract:
In 1988 Whittle introduced an important but intractable class of restless bandit problems which generalise the multiarmed bandit problems of Gittins by allowing state evolution for passive projects. Whittle's account deployed a Lagrangian relaxation of the optimisation problem to develop an index heuristic. Despite a developing body of evidence (both theoretical and empirical) which underscores the strong performance of Whittle's index policy, a continuing challenge to implementation is the need to establish that the competing projects all pass an indexability test. In this paper we employ Gittins' index theory to establish the indexability of (inter alia) general families of restless bandits which arise in problems of machine maintenance and stochastic scheduling problems with switching penalties. We also give formulae for the resulting Whittle indices. Numerical investigations testify to the outstandingly strong performance of the index heuristics concerned.
11
Glazebrook, K. D., D. Ruiz-Hernandez, and C. Kirkbride. "Some indexable families of restless bandit problems." Advances in Applied Probability 38, no. 03 (September 2006): 643–72. http://dx.doi.org/10.1017/s000186780000121x.
Full textAbstract:
In 1988 Whittle introduced an important but intractable class of restless bandit problems which generalise the multiarmed bandit problems of Gittins by allowing state evolution for passive projects. Whittle's account deployed a Lagrangian relaxation of the optimisation problem to develop an index heuristic. Despite a developing body of evidence (both theoretical and empirical) which underscores the strong performance of Whittle's index policy, a continuing challenge to implementation is the need to establish that the competing projects all pass an indexability test. In this paper we employ Gittins' index theory to establish the indexability of (inter alia) general families of restless bandits which arise in problems of machine maintenance and stochastic scheduling problems with switching penalties. We also give formulae for the resulting Whittle indices. Numerical investigations testify to the outstandingly strong performance of the index heuristics concerned.
12
Righter, Rhonda, and J. George Shanthikumar. "Independently Expiring Multiarmed Bandits." Probability in the Engineering and Informational Sciences 12, no. 4 (October 1998): 453–68. http://dx.doi.org/10.1017/s0269964800005325.
Full textAbstract:
We give conditions on the optimality of an index policy for multiarmed bandits when arms expire independently. We also give a new simple proof of the optimality of the Gittins index policy for the classic multiarmed bandit problem.
13
Tsitsiklis, John N. "A Short Proof of the Gittins Index Theorem." Annals of Applied Probability 4, no. 1 (February 1994): 194–99. http://dx.doi.org/10.1214/aoap/1177005207.
Full text
14
Fay, N. A., and J. C. Walrand. "On approximately optimal index strategies for generalised arm problems." Journal of Applied Probability 28, no. 3 (September 1991): 602–12. http://dx.doi.org/10.2307/3214495.
Full textAbstract:
Nash has extended Gittins' work to describe optimal strategies for a class of generalised bandit problems. Here we use a forwards induction argument to analyse ε -optimal strategies for generalised bandit problems. An evaluation procedure for such problems is described; this may be used to analyse models in research planning and stochastic scheduling.
15
Fay, N. A., and J. C. Walrand. "On approximately optimal index strategies for generalised arm problems." Journal of Applied Probability 28, no. 03 (September 1991): 602–12. http://dx.doi.org/10.1017/s0021900200042455.
Full textAbstract:
Nash has extended Gittins' work to describe optimal strategies for a class of generalised bandit problems. Here we use a forwards induction argument to analyse ε -optimal strategies for generalised bandit problems. An evaluation procedure for such problems is described; this may be used to analyse models in research planning and stochastic scheduling.
16
Ding, Zi, and Ilya O. Ryzhov. "Optimal learning with non-Gaussian rewards." Advances in Applied Probability 48, no. 1 (March 2016): 112–36. http://dx.doi.org/10.1017/apr.2015.9.
Full textAbstract:
Abstract We propose a novel theoretical characterization of the optimal 'Gittins index' policy in multi-armed bandit problems with non-Gaussian, infinitely divisible reward distributions. We first construct a continuous-time, conditional Lévy process which probabilistically interpolates the sequence of discrete-time rewards. When the rewards are Gaussian, this approach enables an easy connection to the convenient time-change properties of a Brownian motion. Although no such device is available in general for the non-Gaussian case, we use optimal stopping theory to characterize the value of the optimal policy as the solution to a free-boundary partial integro-differential equation (PIDE). We provide the free-boundary PIDE in explicit form under the specific settings of exponential and Poisson rewards. We also prove continuity and monotonicity properties of the Gittins index in these two problems, and discuss how the PIDE can be solved numerically to find the optimal index value of a given belief state.
17
Glazebrook, K. D., and S. Greatrix. "On transforming an index for generalised bandit problems." Journal of Applied Probability 32, no. 1 (March 1995): 168–82. http://dx.doi.org/10.2307/3214927.
Full textAbstract:
Nash (1980) demonstrated that index policies are optimal for a class of generalised bandit problem. A transform of the index concerned has many of the attributes of the Gittins index. The transformed index is positive-valued, with maximal values yielding optimal actions. It may be characterised as the value of a restart problem and is hence computable via dynamic programming methodologies. The transformed index can also be used in procedures for policy evaluation.
18
Glazebrook, K. D., and S. Greatrix. "On transforming an index for generalised bandit problems." Journal of Applied Probability 32, no. 01 (March 1995): 168–82. http://dx.doi.org/10.1017/s0021900200102633.
Full textAbstract:
Nash (1980) demonstrated that index policies are optimal for a class of generalised bandit problem. A transform of the index concerned has many of the attributes of the Gittins index. The transformed index is positive-valued, with maximal values yielding optimal actions. It may be characterised as the value of a restart problem and is hence computable via dynamic programming methodologies. The transformed index can also be used in procedures for policy evaluation.
19
Aalto, Samuli, Urtzi Ayesta, and Rhonda Righter. "PROPERTIES OF THE GITTINS INDEX WITH APPLICATION TO OPTIMAL SCHEDULING." Probability in the Engineering and Informational Sciences 25, no. 3 (May 17, 2011): 269–88. http://dx.doi.org/10.1017/s0269964811000015.
Full textAbstract:
We consider the optimal scheduling problem for a single-server queue without arrivals. We allow preemptions, and our purpose is to minimize the expected flow time. The optimal nonanticipating discipline is known to be the Gittins index policy, which, however, is defined in an implicit way. Until now, its general behavior in this specific problem has been characterized only in a few special cases. In this article, we give as complete a characterization as possible. It turns out that the optimal policy always belongs to the family of multilevel processor sharing disciplines.
20
Aalto, Samuli, Urtzi Ayesta, and Rhonda Righter. "On the Gittins index in the M/G/1 queue." Queueing Systems 63, no. 1-4 (September 24, 2009): 437–58. http://dx.doi.org/10.1007/s11134-009-9141-x.
Full text
21
Dunn, R. T., and K. D. Glazebrook. "The performance of index-based policies for bandit problems with stochastic machine availability." Advances in Applied Probability 33, no. 2 (June 2001): 365–90. http://dx.doi.org/10.1017/s0001867800010843.
Full textAbstract:
We consider generalisations of two classical stochastic scheduling models, namely the discounted branching bandit and the discounted multi-armed bandit, to the case where the collection of machines available for processing is itself a stochastic process. Under rather mild conditions on the machine availability process we obtain performance guarantees for a range of controls based on Gittins indices. Various forms of asymptotic optimality are established for index-based limit policies as the discount rate approaches 0.
22
Whittle, P. "Restless bandits: activity allocation in a changing world." Journal of Applied Probability 25, A (1988): 287–98. http://dx.doi.org/10.2307/3214163.
Full textAbstract:
We consider a population of n projects which in general continue to evolve whether in operation or not (although by different rules). It is desired to choose the projects in operation at each instant of time so as to maximise the expected rate of reward, under a constraint upon the expected number of projects in operation. The Lagrange multiplier associated with this constraint defines an index which reduces to the Gittins index when projects not being operated are static. If one is constrained to operate m projects exactly then arguments are advanced to support the conjecture that, for m and n large in constant ratio, the policy of operating the m projects of largest current index is nearly optimal. The index is evaluated for some particular projects.
23
Whittle, P. "Restless bandits: activity allocation in a changing world." Journal of Applied Probability 25, A (1988): 287–98. http://dx.doi.org/10.1017/s0021900200040420.
Full textAbstract:
We consider a population of n projects which in general continue to evolve whether in operation or not (although by different rules). It is desired to choose the projects in operation at each instant of time so as to maximise the expected rate of reward, under a constraint upon the expected number of projects in operation. The Lagrange multiplier associated with this constraint defines an index which reduces to the Gittins index when projects not being operated are static. If one is constrained to operate m projects exactly then arguments are advanced to support the conjecture that, for m and n large in constant ratio, the policy of operating the m projects of largest current index is nearly optimal. The index is evaluated for some particular projects.
24
Tan, Cheng, Changbao Xu, Lin Yang, and Wing Shing Wong. "Gittins index based control policy for a class of pursuit-evasion problems." IET Control Theory & Applications 12, no. 1 (January 2, 2018): 110–18. http://dx.doi.org/10.1049/iet-cta.2017.0398.
Full text
25
Sonin, Isaac M. "A generalized Gittins index for a Markov chain and its recursive calculation." Statistics & Probability Letters 78, no. 12 (September 2008): 1526–33. http://dx.doi.org/10.1016/j.spl.2008.01.049.
Full text
26
Rieder, Ulrich, and Jürgen Weishaupt. "Customer Scheduling with Incomplete Information." Probability in the Engineering and Informational Sciences 9, no. 2 (April 1995): 269–84. http://dx.doi.org/10.1017/s0269964800003855.
Full textAbstract:
A stochastic scheduling model with linear waiting costs and unknown routing probabilities is considered. Using a Bayesian approach and methods of Bayesian dynamic programming, we investigate the finite-horizon stochastic scheduling problem with incomplete information. In particular, we study an equivalent nonstationary bandit model and show the monotonicity of the total expected reward and of the Gittins index. We derive the monotonicity and well-known structural properties of the (greatest) maximizers, the so-called stay-on-a-winnerproperty and the stopping-property. The monotonicity results are based on a special partial ordering on .
27
Girlich, H. J., and A. Worku. "Sensitivity of the gittins index in the contiuous time two-armed bandit problem." Optimization 38, no. 4 (January 1996): 367–78. http://dx.doi.org/10.1080/02331939608844264.
Full text
28
Edwards, James, Paul Fearnhead, and Kevin Glazebrook. "ON THE IDENTIFICATION AND MITIGATION OF WEAKNESSES IN THE KNOWLEDGE GRADIENT POLICY FOR MULTI-ARMED BANDITS." Probability in the Engineering and Informational Sciences 31, no. 2 (September 13, 2016): 239–63. http://dx.doi.org/10.1017/s0269964816000279.
Full textAbstract:
The knowledge gradient (KG) policy was originally proposed for online ranking and selection problems but has recently been adapted for use in online decision-making in general and multi-armed bandit problems (MABs) in particular. We study its use in a class of exponential family MABs and identify weaknesses, including a propensity to take actions which are dominated with respect to both exploitation and exploration. We propose variants of KG which avoid such errors. These new policies include an index heuristic, which deploys a KG approach to develop an approximation to the Gittins index. A numerical study shows this policy to perform well over a range of MABs including those for which index policies are not optimal. While KG does not take dominated actions when bandits are Gaussian, it fails to be index consistent and appears not to enjoy a performance advantage over competitor policies when arms are correlated to compensate for its greater computational demands.
29
Talias, Michael A. "Optimal decision indices for R&D project evaluation in the pharmaceutical industry: Pearson index versus Gittins index." European Journal of Operational Research 177, no. 2 (March 2007): 1105–12. http://dx.doi.org/10.1016/j.ejor.2006.01.011.
Full text
30
Glazebrook, K. D., and R. Minty. "A Generalized Gittins Index for a Class of Multiarmed Bandits with General Resource Requirements." Mathematics of Operations Research 34, no. 1 (February 2009): 26–44. http://dx.doi.org/10.1287/moor.1080.0342.
Full text
31
Banks, Jeffrey S., and Rangarajan K. Sundaram. "A class of bandit problems yielding myopic optimal strategies." Journal of Applied Probability 29, no. 3 (September 1992): 625–32. http://dx.doi.org/10.2307/3214899.
Full textAbstract:
We consider the class of bandit problems in which each of the n ≧ 2 independent arms generates rewards according to one of the same two reward distributions, and discounting is geometric over an infinite horizon. We show that the dynamic allocation index of Gittins and Jones (1974) in this context is strictly increasing in the probability that an arm is the better of the two distributions. It follows as an immediate consequence that myopic strategies are the uniquely optimal strategies in this class of bandit problems, regardless of the value of the discount parameter or the shape of the reward distributions. Some implications of this result for bandits with Bernoulli reward distributions are given.
32
Klimmek, Martin. "Parameter Dependent Optimal Thresholds, Indifference Levels and Inverse Optimal Stopping Problems." Journal of Applied Probability 51, no. 2 (June 2014): 492–511. http://dx.doi.org/10.1239/jap/1402578639.
Full textAbstract:
Consider the classic infinite-horizon problem of stopping a one-dimensional diffusion to optimise between running and terminal rewards, and suppose that we are given a parametrised family of such problems. We provide a general theory of parameter dependence in infinite-horizon stopping problems for which threshold strategies are optimal. The crux of the approach is a supermodularity condition which guarantees that the family of problems is indexable by a set-valued map which we call the indifference map. This map is a natural generalisation of the allocation (Gittins) index, a classical quantity in the theory of dynamic allocation. Importantly, the notion of indexability leads to a framework for inverse optimal stopping problems.
33
Klimmek, Martin. "Parameter Dependent Optimal Thresholds, Indifference Levels and Inverse Optimal Stopping Problems." Journal of Applied Probability 51, no. 02 (June 2014): 492–511. http://dx.doi.org/10.1017/s0021900200011384.
Full textAbstract:
Consider the classic infinite-horizon problem of stopping a one-dimensional diffusion to optimise between running and terminal rewards, and suppose that we are given a parametrised family of such problems. We provide a general theory of parameter dependence in infinite-horizon stopping problems for which threshold strategies are optimal. The crux of the approach is a supermodularity condition which guarantees that the family of problems is indexable by a set-valued map which we call the indifference map. This map is a natural generalisation of the allocation (Gittins) index, a classical quantity in the theory of dynamic allocation. Importantly, the notion of indexability leads to a framework for inverse optimal stopping problems.
34
Banks, Jeffrey S., and Rangarajan K. Sundaram. "A class of bandit problems yielding myopic optimal strategies." Journal of Applied Probability 29, no. 03 (September 1992): 625–32. http://dx.doi.org/10.1017/s0021900200043448.
Full textAbstract:
We consider the class of bandit problems in which each of the n ≧ 2 independent arms generates rewards according to one of the same two reward distributions, and discounting is geometric over an infinite horizon. We show that the dynamic allocation index of Gittins and Jones (1974) in this context is strictly increasing in the probability that an arm is the better of the two distributions. It follows as an immediate consequence that myopic strategies are the uniquely optimal strategies in this class of bandit problems, regardless of the value of the discount parameter or the shape of the reward distributions. Some implications of this result for bandits with Bernoulli reward distributions are given.
35
Villar, Sofía S., James Wason, and Jack Bowden. "Response‐adaptive randomization for multi‐arm clinical trials using the forward looking Gittins index rule." Biometrics 71, no. 4 (June 22, 2015): 969–78. http://dx.doi.org/10.1111/biom.12337.
Full text
36
Savelov, M. P. "Gittins Index for Simple Family of Markov Bandit Processes with Switching Cost and No Discounting." Theory of Probability & Its Applications 64, no. 3 (January 2019): 355–64. http://dx.doi.org/10.1137/s0040585x97t989544.
Full text
37
Pandelis, Dimitrios G., and Demosthenis Teneketzis. "On the optimality of the Gittins index rule for multi-armed bandits with multiple plays." Mathematical Methods of Operations Research (ZOR) 50, no. 3 (December 14, 1999): 449–61. http://dx.doi.org/10.1007/s001860050080.
Full text
38
Kallenberg, Lodewijk C. M. "A Note on M. N. Katehakis' and Y.-R. Chen's Computation of the Gittins Index." Mathematics of Operations Research 11, no. 1 (February 1986): 184–86. http://dx.doi.org/10.1287/moor.11.1.184.
Full text
39
Scully, Ziv. "A New Toolbox for Scheduling Theory." ACM SIGMETRICS Performance Evaluation Review 50, no. 3 (December 30, 2022): 3–6. http://dx.doi.org/10.1145/3579342.3579344.
Full textAbstract:
Queueing delays are ubiquitous in many domains, including computer systems, service systems, communication networks, supply chains, and transportation. Queueing and scheduling theory provide a rigorous basis for understanding how to reduce delays with scheduling, including evaluating policy performance and guiding policy design. Unfortunately, stateof- the-art theory fails to address many practical concerns. For example, scheduling theory seldom treats nontrivial preemption limitations, and there is very little theory for scheduling in multiserver queues. My thesis presents two new, broadly applicable tools that greatly expand the reach of scheduling theory, using each to solve multiple open problems. The first tool, called "SOAP", is a new unifying theory of scheduling in single-server queues, specifically the M/G/1 model. SOAP characterizes the delay distribution of a broad space of policies, most of which have never been analyzed before. Such policies include the Gittins index policy, which minimizes mean delay in low-information settings, and many policies with preemption limitations. The second tool, called "WINE", is a new queueing identity that complements Little's law. WINE enables a new method of analyzing complex queueing systems by relating them to simpler systems. This results in the first delay bounds for Shortest Remaining Processing Time (SRPT) and the Gittins policy in multiserver queues, specifically the M/G/k model. This abstract gives a brief overview of my thesis, describing what the SOAP and WINE tools do, the key ideas underlying them, and the open problems they help solve.
40
Whittle, P. "Tax problems in the undiscounted case." Journal of Applied Probability 42, no. 3 (September 2005): 754–65. http://dx.doi.org/10.1239/jap/1127322025.
Full textAbstract:
The aim of this paper is to evaluate the performance of the optimal policy (the Gittins index policy) for open tax problems of the type considered by Klimov in the undiscounted limit. In this limit, the state-dependent part of the cost is linear in the state occupation numbers for the multi-armed bandit, but is quadratic for the tax problem. The discussion of the passage to the limit for the tax problem is believed to be largely new; the principal novelty is our evaluation of the matrix of the quadratic form. These results are confirmed by a dynamic programming analysis, which also suggests how the optimal policy should be modified when resources can be freely deployed only within workstations, rather than system-wide.
41
Whittle, P. "Tax problems in the undiscounted case." Journal of Applied Probability 42, no. 03 (September 2005): 754–65. http://dx.doi.org/10.1017/s0021900200000759.
Full textAbstract:
The aim of this paper is to evaluate the performance of the optimal policy (the Gittins index policy) for open tax problems of the type considered by Klimov in the undiscounted limit. In this limit, the state-dependent part of the cost is linear in the state occupation numbers for the multi-armed bandit, but is quadratic for the tax problem. The discussion of the passage to the limit for the tax problem is believed to be largely new; the principal novelty is our evaluation of the matrix of the quadratic form. These results are confirmed by a dynamic programming analysis, which also suggests how the optimal policy should be modified when resources can be freely deployed only within workstations, rather than system-wide.
42
Niño-Mora, José. "A (2/3)n3Fast-Pivoting Algorithm for the Gittins Index and Optimal Stopping of a Markov Chain." INFORMS Journal on Computing 19, no. 4 (November 2007): 596–606. http://dx.doi.org/10.1287/ijoc.1060.0206.
Full text
43
Smith, Adam L., and Sofía S. Villar. "Bayesian adaptive bandit-based designs using the Gittins index for multi-armed trials with normally distributed endpoints." Journal of Applied Statistics 45, no. 6 (June 28, 2017): 1052–76. http://dx.doi.org/10.1080/02664763.2017.1342780.
Full text
44
Villar, Sofía S., and William F. Rosenberger. "Covariate-adjusted response-adaptive randomization for multi-arm clinical trials using a modified forward looking Gittins index rule." Biometrics 74, no. 1 (July 6, 2017): 49–57. http://dx.doi.org/10.1111/biom.12738.
Full text
45
Esposito Frank, Maria. "Tobias Foster Gittes. Boccaccio's Naked Muse: Eros, Culture, and the Mythopoeic Imagination. Toronto: University of Toronto Press, 2008. xii + 370 pp. index. bibl. $65. ISBN: 978–0–8020–9204–5." Renaissance Quarterly 62, no. 1 (2009): 203–4. http://dx.doi.org/10.1086/598395.
Full text
46
Yao, Yuan, Marco Paolieri, and Leana Golubchik. "Sojourn Time Minimization of Successful Jobs." ACM SIGMETRICS Performance Evaluation Review 50, no. 2 (August 30, 2022): 24–26. http://dx.doi.org/10.1145/3561074.3561083.
Full textAbstract:
Due to a growing interest in deep learning applications [5], compute-intensive and long-running (hours to days) training jobs have become a significant component of datacenter workloads. A large fraction of these jobs is often exploratory, with the goal of determining the best model structure (e.g., the number of layers and channels in a convolutional neural network), hyperparameters (e.g., the learning rate), and data augmentation strategies for the target application. Notably, training jobs are often terminated early if their learning metrics (e.g., training and validation accuracy) are not converging, with only a few completing successfully. For this motivating application, we consider the problem of scheduling a set of jobs that can be terminated at predetermined checkpoints with known probabilities estimated from historical data. We prove that, in order to minimize the time to complete the first K successful jobs on a single server, optimal scheduling does not require preemption (even when preemption overhead is negligible) and provide an optimal policy; advantages of this policy are quantified through simulation. Related Work. While job scheduling has been investigated extensively in many scenarios (see [6] and [2] for a survey of recent result), most policies require that the cost of waiting times of each job be known at scheduling time; in contrast, in our setting the scheduler does not know which job will be the K-th successful job, and sojourn times of subsequent jobs do not contribute to the target metric. For example, [4, 3] minimize makespan (i.e., the time to complete all jobs) for known execution times and waiting time costs; similarly, Gittins index [1] and SR rank [7] minimize expected sojourn time of all jobs, i.e., both successfully completed jobs and jobs terminated early. Unfortunately, scheduling policies not distinguishing between these two types of jobs may favor jobs where the next stage is short and leads to early termination with high probability, which is an undesirable outcome in our applications of interest.
47
Christiansen, S. N., L. Midtbøll Ørnbjerg, S. H. Rasmussen, A. G. Loft, J. K. Wallman, F. Iannone, B. Michelsen, et al. "OP0220 SECULAR TRENDS IN BASELINE CHARACTERISTICS, TREATMENT RETENTION AND RESPONSE RATES IN 17453 BIONAÏVE PSORIATIC ARTHRITIS PATIENTS INITIATING TNFI – RESULTS FROM THE EUROSPA COLLABORATION." Annals of the Rheumatic Diseases 80, Suppl 1 (May 19, 2021): 131.2–132. http://dx.doi.org/10.1136/annrheumdis-2021-eular.422.
Full textAbstract:
Background:Knowledge of changes over time in baseline characteristics and tumor necrosis factor inhibitor (TNFi) response in bionaïve psoriatic arthritis (PsA) patients treated in routine care is limited.Objectives:To investigate secular trends in baseline characteristics and retention, remission and response rates in PsA patients initiating a first TNFi.Methods:Prospectively collected data on bionaïve PsA patients starting TNFi in routine care from 15 European countries were pooled. According to year of TNFi initiation, three groups were defined a priori based on bDMARD availability: Group A (1999–2008), Group B (2009–2014) and Group C (2015–2018).Retention rates (Kaplan-Meier), crude and LUNDEX adjusted1 remission (Disease Activity Score (DAS28) <2.6, 28-joint Disease Activity index for PsA (DAPSA28) ≤4, Clinical Disease Activity Index (CDAI) ≤2.8) and ACR50 response rates were assessed at 6, 12 and 24 months. No statistical comparisons were made.Results:A total of 17453 PsA patients were included (4069, 7551 and 5833 in groups A, B and C).Patients in group A were older and had longer disease duration compared to B and C. Retention rates at 6, 12 and 24 months were highest in group A (88%/77%/64%) but differed little between B (83%/69%/55%) and C (84%/70%/56%).Baseline disease activity was higher in group A than in B and C (DAS28: 4.6/4.3/4.0, DAPSA28: 29.9/25.7/24.0, CDAI: 21.8/20.0/18.6), and this persisted at 6 and 12 months. Crude and LUNDEX adjusted remission rates at 6 and 12 months tended to be lowest in group A, although crude/LUNDEX adjusted ACR50 response rates at all time points were highest in group A. At 24 months, disease activity and remission rates were similar in the three groups (Table).Table 1.Secular trends in baseline characteristics, treatment retention, remission and response rates in European PsA patients initiating a 1st TNFiBaseline characteristicsGroup A(1999–2008)Group B(2009–2014)Group C(2015–2018)Age, median (IQR)62 (54–72)58 (49–67)54 (45–62)Male, %514847Years since diagnosis, median (IQR)5 (2–10)3 (1–9)3 (1–8)Smokers, %161717DAS28, median (IQR)4.6 (3.7–5.3)4.3 (3.4–5.1)4.0 (3.2–4.8)DAPSA28, median (IQR)29.9 (19.3–41.8)25.7 (17.2–38.1)24.0 (16.1–35.5)CDAI, median (IQR)21.8 (14.0–31.1)20.0 (13.0–29.0)18.6 (12.7–26.1)TNFi drug, % (Adalimumab / Etanercept / Infliximab / Certolizumab / Golimumab)27 / 43 / 30 / 0 / 036 / 31 / 14 / 5 / 1421 / 40 / 21 / 8 / 10Follow up6 months12 months24 monthsGr AGr BGr CGr AGr BGr CGr AGr BGr CRetention rates, % (95% CI)88 (87–89)83 (82–84)84 (83–85)79 (78–80)72 (71–73)72 (71–73)68 (67–69)60 (59–61)60 (59–62)DAS28, median (IQR)2.7 (1.9–3.6)2.4 (1.7–3.4)2.3 (1.7–3.2)2.5 (1.8–3.4)2.2 (1.6–3.1)2.1 (1.6–2.9)2.1 (1.6–3.1)2.0 (1.6–2.9)1.9 (1.5–2.6)DAPSA28, median (IQR)10.6 (4.8–20.0)9.5 (3.9–18.3)8.7 (3.6–15.9)9.1 (4.1–17.8)7.7 (3.1–15.4)7.6 (2.9–14.4)6.7 (2.7–13.7)6.6 (2.7–13.5)5.9 (2.4–11.8)CDAI, median (IQR)7.8 (3.0–15.2)8.0 (3.0–15.0)6.4 (2.6–12.2)6.4 (2.5–13.0)6.2 (2.5–12.1)5.8 (2.2–11.4)5.0 (2.0–11.0)5.5 (2.0–11.2)5.0 (2.0–9.0)DAS28 remission, %, c/L47 / 4255 / 4661 / 5153 / 4362 / 4566 / 4864 / 4268 / 3775 / 41DAPSA28 remission, %, c/L22 / 1926 / 2228 / 2325 / 2031 / 2232 / 2336 / 2334 / 1938 / 21CDAI remission, %, c/L23 / 2123 / 1926 / 2227 / 2127 / 2029 / 2134 / 2231 / 1735 / 19ACR50 response, %, c/L26 / 2322 / 1824 / 2027 / 2223 / 1721 / 1523 / 1518 / 1014 / 8Gr, Group; c/L, crude/LUNDEX.Conclusion:Over the past 20 years, patient age, disease duration and disease activity level at the start of the first TNFi in PsA patients have decreased. Furthermore, TNFi retention rates have decreased while remission rates have increased, especially remission rates within the first year of treatment. These findings may reflect a greater awareness of early diagnosis in PsA patients, a lowered threshold for initiating TNFi and the possibility for earlier switching in patients with inadequate treatment response.References:[1]Arthritis Rheum 2006; 54: 600-6.Acknowledgements:Novartis Pharma AG and IQVIA for supporting the EuroSpA Research Collaboration Network.Disclosure of Interests:Sara Nysom Christiansen Speakers bureau: BMS and GE, Grant/research support from: Novartis, Lykke Midtbøll Ørnbjerg Grant/research support from: Novartis, Simon Horskjær Rasmussen: None declared, Anne Gitte Loft Speakers bureau: AbbVie, Janssen, Lilly, MSD, Novartis, Pfizer, UCB, Consultant of: AbbVie, Janssen, Lilly, MSD, Novartis, Pfizer, UCB, Grant/research support from: Novartis, Johan K Wallman Consultant of: Celgene, Eli Lilly, Novartis, Florenzo Iannone Speakers bureau: Abbvie, MSD, Novartis, Pfizer and BMS, Brigitte Michelsen Consultant of: Novartis, Grant/research support from: Novartis, Michael J. Nissen Speakers bureau: Novartis, Eli Lilly, Celgene, and Pfizer, Consultant of: Novartis, Eli Lilly, Celgene, and Pfizer, Jakub Zavada: None declared, Maria Jose Santos Speakers bureau: AbbVie, Novartis, Pfizer, Manuel Pombo-Suarez: None declared, Kari Eklund: None declared, Matija Tomsic Speakers bureau: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Consultant of: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Björn Gudbjornsson Speakers bureau: Amgen and Novartis, İsmail Sari: None declared, Catalin Codreanu Speakers bureau: AbbVie, Amgen, Egis, Novartis, Pfizer, UCB, Grant/research support from: AbbVie, Amgen, Egis, Novartis, Pfizer, UCB, Daniela Di Giuseppe: None declared, Bente Glintborg Grant/research support from: Pfizer, Biogen, AbbVie, Marco Sebastiani: None declared, Karen Minde Fagerli: None declared, Burkhard Moeller: None declared, Karel Pavelka Speakers bureau: AbbVie, Roche, MSD, UCB, Pfizer, Novartis, Egis, Gilead, Eli Lilly, Consultant of: AbbVie, Roche, MSD, UCB, Pfizer, Novartis, Egis, Gilead, Eli Lilly, Anabela Barcelos: None declared, Carlos Sánchez-Piedra: None declared, Heikki Relas: None declared, Ziga Rotar Speakers bureau: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Consultant of: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Thorvardur Love: None declared, Servet Akar: None declared, Ruxandra Ionescu Speakers bureau: Abbvie, Amgen, Boehringer-Ingelheim Eli-Lilly,Novartis, Pfizer, Sandoz, UCB, Gary Macfarlane Grant/research support from: GlaxoSmithKline, Marleen G.H. van de Sande: None declared, Merete L. Hetland Speakers bureau: Abbvie, Biogen, BMS, Celltrion, Eli Lilly, Janssen Biologics B.V, Lundbeck Fonden, MSD, Pfizer, Roche, Samsung Biopies, Sandoz, Novartis., Mikkel Østergaard Speakers bureau: AbbVie, BMS, Boehringer-Ingelheim, Celgene, Eli-Lilly, Centocor, GSK, Hospira, Janssen, Merck, Mundipharma, Novartis, Novo, Orion, Pfizer, Regeneron, Schering-Plough, Roche, Takeda, UCB and Wyeth, Consultant of: AbbVie, BMS, Boehringer-Ingelheim, Celgene, Eli-Lilly, Centocor, GSK, Hospira, Janssen, Merck, Mundipharma, Novartis, Novo, Orion, Pfizer, Regeneron, Schering-Plough, Roche, Takeda, UCB and Wyeth
48
Midtbøll Ørnbjerg, L., S. N. Christiansen, S. H. Rasmussen, A. G. Loft, U. Lindström, J. Zavada, F. Iannone, et al. "POS0027 SECULAR TRENDS IN BASELINE CHARACTERISTICS, TREATMENT RETENTION AND RESPONSE RATES IN 27189 BIO-NAÏVE AXIAL SPONDYLOARTHRITIS PATIENTS INITIATING TNFI – RESULTS FROM THE EUROSPA COLLABORATION." Annals of the Rheumatic Diseases 80, Suppl 1 (May 19, 2021): 217–18. http://dx.doi.org/10.1136/annrheumdis-2021-eular.589.
Full textAbstract:
Background:Knowledge of changes over time in baseline characteristics and tumor necrosis factor inhibitor (TNFi) response in bio-naïve axial spondyloarthritis (axSpA) patients treated in routine care is limited.Objectives:To investigate secular trends in baseline characteristics and retention, remission and response rates in axSpA patients initiating a first TNFi.Methods:Prospectively collected data on bio-naïve axSpA patients starting TNFi in routine care from 15 European countries were pooled. According to year of TNFi initiation, three groups were defined a priori based on bDMARD availability: Group A (1999–2008), Group B (2009–2014) and Group C (2015–2018). Retention rates (Kaplan-Meier), crude and LUNDEX adjusted1 remission (Ankylosing Spondylitis Disease Activity Score (ASDAS) <1.3, Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) <20) and response (ASDAS Major and Clinically Important Improvement (MI/CII), BASDAI 50) rates were assessed at 6, 12 and 24 months. No statistical comparisons were made.Results:In total, 27189 axSpA patients were included (5945, 11255 and 9989 in groups A, B and C).At baseline, patients in group A were older, had longer disease duration and a larger proportion of male and HLA-B27 positive patients compared to B and C, whereas disease activity was similar across groups.Retention rates at 6, 12 and 24 months were highest in group A (88%/81%/71%) but differed little between B (84%/74%/64%) and C (85%/76%/67%).In all groups, median ASDAS and BASDAI had decreased markedly at 6 months (Table 1). The ASDAS values at 12 and 24 months and BASDAI at 24 months were higher in group A compared with groups B and C. Similarly, crude remission and response rates were lowest in group A. After adjustments for drug retention (LUNDEX), remission and response rates showed less pronounced between-group differences regarding ASDAS measures and no relevant differences regarding BASDAI measures.Conclusion:Nowadays, axSpA patients initiating TNFi are younger with shorter disease duration and more frequently female and HLA-B27 negative than previously, while baseline disease activity is unchanged. Drug retention rates have decreased, whereas crude remission and response rates have increased. This may indicate expanded indication but also a stable disease activity threshold for TNFi initiation over time, an increased focus on targeting disease remission and more available treatment options.References:[1]Arthritis Rheum 2006; 54: 600-6.Table 1.Secular trends in baseline characteristics, treatment retention, remission and response rates in European axSpA patients initiating a 1st TNFiBaseline characteristicsGroup A(1999–2008)Group B(2009–2014)Group C(2015–2018)Age, years, median (IQR)57 (49–66)51 (42–60)46 (37–56)Male, %666057HLA-B27, %877772Years since diagnosis, median (IQR)5 (1–12)2 (0–8)2 (0–7)Smokers, %232425ASDAS, median (IQR)3.5 (2.8–4.1)3.4 (2.8–4.1)3.5 (2.8–4.1)BASDAI, median, (IQR)57 (42–71)59 (43–72)57 (41–71)TNFi drug, % (Adalimumab /Etanercept / Infliximab /Certolizumab / Golimumab)22 / 35 / 43 / 0 / 037 / 21 / 20 / 4 / 1827 / 28 / 24 / 8 / 13Follow up6 months12 months24 monthsGr AGr BGr CGr AGr BGr CGr AGr BGr CRetention rates, %, (95% CI)88 (88–89)84 (83–85)85 (84–86)81 (80–82)74 (74–75)76 (75–76)71 (70–72)64 (63–65)67 (66–68)ASDAS, median, (IQR)1.8 (1.2–2.8)1.9 (1.2–2.8)1.8 (1.2–2.6)1.9 (1.3–2.6)1.7 (1.2–2.5)1.6 (1.1–2.4)1.9 (1.4–2.6)1.7 (1.1–2.4)1.5 (1.1–2.2)ASDAS inactive disease, %, c/L28 / 2528 / 2430 / 2624 / 1932 / 2434 / 2623 / 1634 / 2039 / 23ASDAS CII, %, c/L57 / 5159 / 5063 / 5461 / 5063 / 4767 / 5159 / 4168 / 4074 / 45ASDAS MI, %, c/L31 / 2732 / 2737 / 3232 / 2637 / 2741 / 3130 / 2042 / 2546 / 28BASDAI, median, (IQR)23 (10–40)26 (11–48)24 (10–44)21 (10–38)23 (10–42)20 (8–39)22 (9–40)20 (8–39)16 (6–35)BASDAI remission, %, c/L44 / 4040 / 3443 / 3645 / 3645 / 3450 / 3844 / 3048 / 2956 / 34BASDAI 50 response, %, c/L53 / 4750 / 4253 / 4557 / 4656 / 4258 / 4457 / 3960 / 3563 / 38Gr, Group; c/L, crude/LUNDEX adjusted.Acknowledgements:Novartis Pharma AG and IQVIA for supporting the EuroSpA Research Collaboration Network.Disclosure of Interests:Lykke Midtbøll Ørnbjerg Grant/research support from: Novartis, Sara Nysom Christiansen Speakers bureau: BMS and GE, Grant/research support from: Novartis, Simon Horskjær Rasmussen: None declared, Anne Gitte Loft Speakers bureau: AbbVie, Janssen, Lilly, MSD, Novartis, Pfizer, UCB, Consultant of: AbbVie, Janssen, Lilly, MSD, Novartis, Pfizer, UCB, Grant/research support from: Novartis, Ulf Lindström: None declared, Jakub Zavada: None declared, Florenzo Iannone: None declared, Fatos Onen: None declared, Michael J. Nissen Speakers bureau: Novartis, Eli Lilly, Celgene, and Pfizer, Consultant of: Novartis, Eli Lilly, Celgene, and Pfizer, Brigitte Michelsen Consultant of: Novartis, Grant/research support from: Novartis, Maria Jose Santos Speakers bureau: AbbVie, Novartis, Pfizer, Gary Macfarlane Grant/research support from: GlaxoSmithKline, Dan Nordström Consultant of: Abbvie, BMS, MSD, Novartis, Pfizer, Roche, UCB, Manuel Pombo-Suarez: None declared, Catalin Codreanu Speakers bureau: AbbVie, Amgen, Egis, Novartis, Pfizer, UCB, Grant/research support from: AbbVie, Amgen, Egis, Novartis, Pfizer, UCB, Matija Tomsic Speakers bureau: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Consultant of: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Irene van der Horst-Bruinsma Speakers bureau: Abbvie, BMS, MSD, Novartis, Pfizer, Lilly, UCB, Björn Gudbjornsson Speakers bureau: Amgen and Novartis, Johan Askling: None declared, Bente Glintborg Grant/research support from: Pfizer, Biogen, AbbVie, Karel Pavelka Speakers bureau: AbbVie, Roche, MSD, UCB, Pfizer, Novartis, Egis, Gilead, Eli Lilly, Consultant of: AbbVie, Roche, MSD, UCB, Pfizer, Novartis, Egis, Gilead, Eli Lilly, Elisa Gremese: None declared, Nurullah Akkoc: None declared, Adrian Ciurea Speakers bureau: Abbvie, Eli-Lilly, MSD, Novartis, Pfizer, Eirik kristianslund: None declared, Anabela Barcelos: None declared, Gareth T. Jones Grant/research support from: Pfizer, AbbVie, UCB, Celgene, Amgen, GSK, Anna-Mari Hokkanen Grant/research support from: MSD, Carlos Sánchez-Piedra: None declared, Ruxandra Ionescu Speakers bureau: Abbvie, Amgen, Boehringer-Ingelheim Eli-Lilly,Novartis, Pfizer, Sandoz, UCB, Ziga Rotar Speakers bureau: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Consultant of: Abbvie, Amgen, Biogen, Medis, MSD, Novartis, Pfizer, Marleen G.H. van de Sande: None declared, Arni Jon Geirsson: None declared, Mikkel Østergaard Speakers bureau: AbbVie, BMS, Boehringer-Ingelheim, Celgene, Eli-Lilly, Centocor, GSK, Hospira, Janssen, Merck, Mundipharma, Novartis, Novo, Orion, Pfizer, Regeneron, Schering-Plough, Roche, Takeda, UCB and Wyeth, Consultant of: AbbVie, BMS, Boehringer-Ingelheim, Celgene, Eli-Lilly, Centocor, GSK, Hospira, Janssen, Merck, Mundipharma, Novartis, Novo, Orion, Pfizer, Regeneron, Schering-Plough, Roche, Takeda, UCB and Wyeth, Merete L. Hetland Speakers bureau: Abbvie, Biogen, BMS, Celltrion, Eli Lilly, Janssen Biologics B.V, Lundbeck Fonden, MSD, Pfizer, Roche, Samsung Biopies, Sandoz, Novartis.
49
Georgiadis, S., M. Riek, C. Polysopoulos, A. Scherer, D. DI Giuseppe, G. T. Jones, M. L. Hetland, et al. "POS0001 CAN SINGLE IMPUTATION TECHNIQUES FOR BASDAI COMPONENTS RELIABLY CALCULATE THE COMPOSITE SCORE IN AXIAL SPONDYLOARTHRITIS PATIENTS?" Annals of the Rheumatic Diseases 81, Suppl 1 (May 23, 2022): 212–13. http://dx.doi.org/10.1136/annrheumdis-2022-eular.1562.
Full textAbstract:
BackgroundIn axial spondyloarthritis (axSpA), Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) is a key patient-reported outcome. However, one or more of its components may be missing when recorded in clinical practice.ObjectivesTo determine whether an individual patient’s BASDAI at a given timepoint can be reliably calculated with different single imputation techniques and to explore the impact of the number of missing components and/or differences between missingness of individual components.MethodsReal-life data from axSpA patients receiving tumour necrosis factor inhibitors (TNFi) from 13 countries in the European Spondyloarthritis (EuroSpA) Research Collaboration Network were utilized [1]. We studied missingness in BASDAI components based on simulations in a complete dataset, where we applied and expanded the approach of Ramiro et al. [2]. After introducing one or more missing components completely at random, BASDAI was calculated from the available components and with three different single imputation techniques: possible middle value (i.e. 50) of the component and mean and median of the available components. Differences between the observed (original) and calculated scores were assessed and correct classification of patients as having BASDAI<40 mm was additionally evaluated. For the setting with one missing component, differences arising between missing one of components 1-4 versus 5-6 were explored. Finally, the performance of imputations in relation to the values of the original score was investigated.ResultsA total of 19,894 axSpA patients with at least one complete BASDAI registration at any timepoint were included. 59,126 complete BASDAI registrations were utilized for the analyses with a mean BASDAI of 38.5 (standard deviation 25.9). Calculating BASDAI from the available components and imputing with mean or median showed similar levels of agreement (Table 1). When allowing one missing component, >90% had a difference of ≤6.9 mm between the original and calculated scores and >95% were correctly classified as BASDAI<40 (Table 1). However, separate analyses of components 1-4 and 5-6 as a function of the BASDAI score suggested that imputing any one of the first four BASDAI components resulted in a level of agreement <90% for specific BASDAI values while imputing one of the stiffness components 5-6 always reached a level of agreement >90% (Figure 1, upper panels). As expected, it was observed that regardless of the BASDAI component set to missing and the imputation technique used, correct classification of patients as BASDAI<40 was less than 95% for values around the cutoff (Figure 1, lower panels).Table 1.Level of agreement between the original and calculated BASDAI and correct classification for BASDAI<40 mmLevel of agreement with Dif≤6.9 mm* (%)Correct classification for BASDAI<40 mm** (%)1 missing componentAvailable93.996.9Value 5073.996.3Mean94.296.8Median93.196.82 missing componentsAvailable83.794.8Value 5040.792.8Mean83.594.8Median82.894.73 missing componentsAvailable71.992.6Value 5028.187.3Mean72.292.6Median69.792.2* The levels of agreement with a difference (Dif) of ≤6.9 mm between the original and calculated scores were based on the half of the smallest detectable change. Agreement of >90% was considered as acceptable. ** Correct classification of >95% was considered as acceptable.Figure 1.Level of agreement between the original and calculated BASDAI and correct classification for BASDAI<40 mm as a function of the original scoreConclusionBASDAI calculation with available components gave similar results to single imputation of missing components with mean or median. Only when missing one of BASDAI components 5 or 6, single imputation techniques can reliably calculate individual BASDAI scores. However, missing any single component value results in misclassification of patients with original BASDAI scores close to 40.References[1]Ørnbjerg et al. (2019). Ann Rheum Dis, 78(11), 1536-1544.[2]Ramiro et al. (2014). Rheumatology, 53(2), 374-376.AcknowledgementsNovartis Pharma AG and IQVIA for supporting the EuroSpA collaboration.Disclosure of InterestsStylianos Georgiadis Grant/research support from: Novartis, Myriam Riek Grant/research support from: Novartis, Christos Polysopoulos Grant/research support from: Novartis, Almut Scherer Grant/research support from: Novartis, Daniela Di Giuseppe: None declared, Gareth T. Jones Speakers bureau: Janssen, Grant/research support from: AbbVie, Pfizer, UCB, Amgen, GSK, Merete Lund Hetland Grant/research support from: Abbvie, Biogen, BMS, Celltrion, Eli Lilly, Janssen Biologics B.V, Lundbeck Fonden, MSD, Medac, Pfizer, Roche, Samsung Biopies, Sandoz, Novartis, Mikkel Østergaard Speakers bureau: Abbvie, BMS, Boehringer-Ingelheim, Celgene, Eli-Lilly, Hospira, Janssen, Merck, Novartis, Novo, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, UCB, Consultant of: Abbvie, BMS, Boehringer-Ingelheim, Celgene, Eli-Lilly, Hospira, Janssen, Merck, Novartis, Novo, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, UCB, Grant/research support from: Abbvie, BMS, Merck, Celgene, Novartis, Simon Horskjær Rasmussen Grant/research support from: Novartis, Johan K Wallman Consultant of: AbbVie, Amgen, Celgene, Eli Lilly, Novartis, Bente Glintborg Grant/research support from: Pfizer, Abbvie, BMS, Anne Gitte Loft Speakers bureau: AbbVie, Janssen, Lilly, MSD, Novartis, Pfizer, Roche, UCB, Consultant of: AbbVie, Janssen, Lilly, MSD, Novartis, Pfizer, Roche, UCB, Karel Pavelka Speakers bureau: Pfizer, MSD, BMS, UCB, Amgen, Egis, Roche, AbbVie, Consultant of: Pfizer, MSD, BMS, UCB, Amgen, Egis, Roche, AbbVie, Jakub Zavada Speakers bureau: Abbvie, Elli-Lilly, Sandoz, Novartis, Egis, UCB, Consultant of: Abbvie, Elli-Lilly, Sandoz, Novartis, Egis, UCB, Merih Birlik: None declared, Ayten Yazici Grant/research support from: Roche, Brigitte Michelsen Grant/research support from: Novartis, Eirik kristianslund: None declared, Adrian Ciurea Speakers bureau: AbbVie, Eli Lilly, Merck Sharp & Dohme, Novartis, Pfizer, Consultant of: AbbVie, Eli Lilly, Merck Sharp & Dohme, Novartis, Pfizer, Michael J. Nissen Speakers bureau: AbbVie, Eli Lilly, Janssens, Novartis, Pfizer, Consultant of: AbbVie, Eli Lilly, Janssens, Novartis, Pfizer, Ana Maria Rodrigues Speakers bureau: Abbvie, Amgen, Consultant of: Abbvie, Amgen, Grant/research support from: Novartis, Pfizer, Amgen, Maria Jose Santos Speakers bureau: Abbvie, AstraZeneca, Lilly, Novartis, Pfizer, Gary Macfarlane Grant/research support from: GSK, Anna-Mari Hokkanen Grant/research support from: MSD, Heikki Relas Speakers bureau: Abbvie, Celgene, Pfizer, UCB, Viatris, Consultant of: Abbvie, Celgene, Pfizer, UCB, Viatris, Catalin Codreanu Speakers bureau: AbbVie, Amgen, Boehringer Ingelheim, Ewopharma, Lilly, Novartis, Pfizer, Consultant of: AbbVie, Amgen, Boehringer Ingelheim, Ewopharma, Lilly, Novartis, Pfizer, Corina Mogosan: None declared, Ziga Rotar Speakers bureau: Abbvie, Novartis, MSD, Medis, Biogen, Eli Lilly, Pfizer, Sanofi, Lek, Janssen, Consultant of: Abbvie, Novartis, MSD, Medis, Biogen, Eli Lilly, Pfizer, Sanofi, Lek, Janssen, Matija Tomsic Speakers bureau: Abbvie, Amgen, Biogen, Eli Lilly, Janssen, Medis, MSD, Novartis, Pfizer, Sanofi, Sandoz-Lek, Consultant of: Abbvie, Amgen, Biogen, Eli Lilly, Janssen, Medis, MSD, Novartis, Pfizer, Sanofi, Sandoz-Lek, Björn Gudbjornsson Speakers bureau: Amgen, Novartis, Consultant of: Amgen, Novartis, Arni Jon Geirsson: None declared, Pasoon Hellamand Grant/research support from: Novartis, Marleen G.H. van de Sande Speakers bureau: Eli Lilly, Novartis, UCB, Janssen, Abbvie, Consultant of: Eli Lilly, Novartis, UCB, Janssen, Abbvie, Grant/research support from: Eli Lilly, Novartis, UCB, Janssen, Abbvie, Isabel Castrejon: None declared, Manuel Pombo-Suarez Consultant of: Abbvie, MSD, Roche, Bruno Frediani: None declared, Florenzo Iannone Speakers bureau: Abbvie, Amgen, AstraZeneca, BMS, Galapagos, Janssen, Lilly, MSD, Novartis, Pfizer, Roche, UCB, Consultant of: Abbvie, Amgen, AstraZeneca, BMS, Galapagos, Janssen, Lilly, MSD, Novartis, Pfizer, Roche, UCB, Lykke Midtbøll Ørnbjerg Grant/research support from: Novartis
50
Michelsen, B., S. Georgiadis, D. DI Giuseppe, A. G. Loft, M. Nissen, F. Iannone, M. Pombo-Suarez, et al. "SAT0430 SECUKINUMAB EFFECTIVENESS IN 1543 PATIENTS WITH PSORIATIC ARTHRITIS TREATED IN ROUTINE CLINICAL PRACTICE IN 13 EUROPEAN COUNTRIES IN THE EuroSpA RESEARCH COLLABORATION NETWORK." Annals of the Rheumatic Diseases 79, Suppl 1 (June 2020): 1169.2–1171. http://dx.doi.org/10.1136/annrheumdis-2020-eular.1413.
Full textAbstract:
Background:There is a lack of real-life evidence on secukinumab effectiveness in psoriatic arthritis (PsA) patients.Objectives:To assess the real-life 6- and 12-month secukinumab retention rates and proportions of patients in remission/low disease activity (LDA) overall, and by prior biologic disease-modifying anti-rheumatic drug (bDMARD)/targeted synthetic (ts)DMARD use.Methods:Data from PsA patients treated with secukinumab in routine care from 13 countries in the European Spondyloarthritis (EuroSpA) Research Collaboration Network were pooled. Patients started secukinumab ≥12 months before date of datacut. Crude and LUNDEX adjusted (crude value adjusted for drug retention) 28-joint Disease Activity index for PSoriatic Arthritis (DAPSA28) and 28-joint Disease Activity Score with CRP (DAS28CRP) remission and LDA rates were calculated. Group comparisons between b/tsDMARD naïve, 1 prior and ≥2 prior b/tsDMARD users were done with ANOVA, Kruskal-Wallis, Chi-square or Kaplan-Meier analyses with log-rank test, as appropriate.Results:A total of 1543 PsA patients were included (Table 1). b/tsDMARD naïve patients had shorter time since diagnosis, higher baseline disease activity, a higher proportion were men and a higher proportion achieved remission. Overall 6/12-month secukinumab retention rates were 86%/74% and significantly higher in b/tsDMARD naïve patients at 12, but not 6 months (Table 2, Figure). Overall, crude 6- and 12-month DAPSA28≤4/DAS28CRP<2.6 were achieved by 13%/34% and 11%/39% of the patients, respectively.Table 1.All patients (n=1543)b/tsDMARD naïve (n=287)1 prior b/tsDMARD (n=333)≥2 prior b/tsDMARDs (n=923)p *Age (years), mean (SD)52 (11)49 (12.3)51 (11)53 (11)<0.001Male, %42%49%46%39%0.003Years since diagnosis, mean (SD)9 (8)7 (8)8 (7)10 (8)<0.001Current smokers, %19%21%22%18%0.23CRP (mg/L), median (IQR)5 (2-12)7 (2-19)4 (2-8)5 (2-11)<0.001DAPSA28, median (IQR)26 (18-37)28 (19-38)22 (13-32)27 (19-38)<0.001DAS28CRP, median (IQR)4.2 (3.3-5.0)4.4 (3.5-5.2)3.8 (2.6-4.5)4.2 (3.4-5.0)<0.001*Comparisons across number of prior b/tsDMARD were done with ANOVA, Kruskal-Wallis or Chi-square test, as appropriateTable 2.MonthsAll patients (n=1543)b/tsDMARD naïve (n=287)1 prior b/tsDMARD (n=333)≥2 prior b/tsDMARDs (n=923)p *Secukinumab retention rate, % (95%CI)686% (84-87%)89% (86-93%)85% (81-89%)85% (82-87%)0.111274% (72-76%)81% (76-86%)76% (71-80%)72% (69-75%)0.006DAPSA28≤4 Crude613%25%11%11%<0.001 LUNDEX11%22%9%9%<0.001 Crude1211%22%11%8%<0.001 LUNDEX7%17%7%5%0.001DAS28CRP<2.6 Crude634%51%33%30%<0.001 LUNDEX29%45%27%24%<0.001 Crude1239%55%41%34%<0.001 LUNDEX26%41%27%21%<0.001DAPSA28 >4 and ≤14 Crude633%42%32%30%0.04 LUNDEX27%37%27%25%0.02 Crude1235%48%36%32%0.009 LUNDEX24%36%24%20%0.004DAS28CRP ≤3.2 Crude652%69%53%47%<0.001 LUNDEX43%61%45%38%<0.001 Crude1255%72%55%50%<0.001 LUNDEX37%54%37%32%<0.001*Comparisons across number of prior b/tsDMARDs were done with Kaplan-Meier with log-rank test or Chi-Square test, as appropriateConclusion:In this real-life study of 1543 patients with PsA in 13 European countries 12-month secukinumab retention was high, and significantly higher for b/tsDMARD naïve patients. Overall, a higher proportion of bionaïve than previous b/tsDMARD users achieved remission, regardless of remission criteria.Acknowledgments:Novartis and IQVIA for supporting the EuroSpA RCNDisclosure of Interests:Brigitte Michelsen Grant/research support from: Research support from Novartis, Consultant of: Consulting fees Novartis, Stylianos Georgiadis Grant/research support from: Novartis, Daniela Di Giuseppe: None declared, Anne Gitte Loft Grant/research support from: Novartis, Consultant of: AbbVie, MSD, Novartis, Pfizer and UCB, Speakers bureau: AbbVie, MSD, Novartis, Pfizer and UCB, Michael Nissen Grant/research support from: Abbvie, Consultant of: Novartis, Lilly, Abbvie, Celgene and Pfizer, Speakers bureau: Novartis, Lilly, Abbvie, Celgene and Pfizer, Florenzo Iannone Consultant of: Speaker and consulting fees from AbbVie, Eli Lilly, Novartis, Pfizer, Roche, Sanofi, UCB, MSD, Speakers bureau: Speaker and consulting fees from AbbVie, Eli Lilly, Novartis, Pfizer, Roche, Sanofi, UCB, MSD, Manuel Pombo-Suarez Consultant of: Janssen, Lilly, MSD and Sanofi., Speakers bureau: Janssen, Lilly, MSD and Sanofi., Heřman Mann: None declared, Ziga Rotar Consultant of: Speaker and consulting fees from Abbvie, Amgen, Biogen, Eli Lilly, Medis, MSD, Novartis, Pfizer, Roche, Sanofi., Speakers bureau: Speaker and consulting fees from Abbvie, Amgen, Biogen, Eli Lilly, Medis, MSD, Novartis, Pfizer, Roche, Sanofi., Kari Eklund Consultant of: Celgene, Lilly, Speakers bureau: Pfizer, Roche, Tore K. Kvien Grant/research support from: Received grants from Abbvie, Hospira/Pfizer, MSD and Roche (not relevant for this abstract)., Consultant of: Have received personal fees from Abbvie, Biogen, BMS, Celltrion, Eli Lily, Hospira/Pfizer, MSD, Novartis, Orion Pharma, Roche, Sandoz, UCB, Sanofi and Mylan (not relevant for this abstract)., Paid instructor for: Have received personal fees from Abbvie, Biogen, BMS, Celltrion, Eli Lily, Hospira/Pfizer, MSD, Novartis, Orion Pharma, Roche, Sandoz, UCB, Sanofi and Mylan (not relevant for this abstract)., Speakers bureau: Have received personal fees from Abbvie, Biogen, BMS, Celltrion, Eli Lily, Hospira/Pfizer, MSD, Novartis, Orion Pharma, Roche, Sandoz, UCB, Sanofi and Mylan (not relevant for this abstract)., Maria Jose Santos Speakers bureau: Novartis and Pfizer, Björn Gudbjornsson Speakers bureau: Novartis and Amgen, Catalin Codreanu Consultant of: Speaker and consulting fees from AbbVie, Accord Healthcare, Alfasigma, Egis, Eli Lilly, Ewopharma, Genesis, Mylan, Novartis, Pfizer, Roche, Sandoz, UCB, Speakers bureau: Speaker and consulting fees from AbbVie, Accord Healthcare, Alfasigma, Egis, Eli Lilly, Ewopharma, Genesis, Mylan, Novartis, Pfizer, Roche, Sandoz, UCB, Sema Yilmaz: None declared, Johan K Wallman Consultant of: AbbVie, Celgene, Eli Lilly, Novartis and UCB Pharma, Cecilie Heegaard Brahe Grant/research support from: Novartis, Burkhard Moeller: None declared, Ennio Giulio Favalli Consultant of: Consultant and/or speaker for BMS, Eli-Lilly, MSD, UCB, Pfizer, Sanofi-Genzyme, Novartis, and Abbvie, Speakers bureau: Consultant and/or speaker for BMS, Eli-Lilly, MSD, UCB, Pfizer, Sanofi-Genzyme, Novartis, and Abbvie, Carlos Sánchez-Piedra: None declared, Lucie Nekvindova: None declared, Matija Tomsic: None declared, Nina Trokovic: None declared, Eirik kristianslund: None declared, Helena Santos Speakers bureau: AbbVie, Eli-Lilly, Janssen, Pfizer, Novartis, Thorvardur Love: None declared, Ruxandra Ionescu Consultant of: Consulting fees from Abbvie, Eli-Lilly, Novartis, Pfizer, Roche, Sandoz, Speakers bureau: Consulting and speaker fees from Abbvie, Eli-Lilly, Novartis, Pfizer, Roche, Sandoz, Yavuz Pehlivan: None declared, Gareth T. Jones Grant/research support from: Pfizer, AbbVie, UCB, Celgene and GSK., Irene van der Horst-Bruinsma Grant/research support from: AbbVie, Novartis, Eli Lilly, Bristol-Myers Squibb, MSD, Pfizer, UCB Pharma, Consultant of: AbbVie, Novartis, Eli Lilly, Bristol-Myers Squibb, MSD, Pfizer, UCB Pharma, Lykke Midtbøll Ørnbjerg Grant/research support from: Novartis, Mikkel Ǿstergaard Grant/research support from: AbbVie, Bristol-Myers Squibb, Celgene, Merck, and Novartis, Consultant of: AbbVie, Bristol-Myers Squibb, Boehringer Ingelheim, Celgene, Eli Lilly, Hospira, Janssen, Merck, Novartis, Novo Nordisk, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, and UCB, Speakers bureau: AbbVie, Bristol-Myers Squibb, Boehringer Ingelheim, Celgene, Eli Lilly, Hospira, Janssen, Merck, Novartis, Novo Nordisk, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, and UCB, Merete L. Hetland Grant/research support from: BMS, MSD, AbbVie, Roche, Novartis, Biogen and Pfizer, Consultant of: Eli Lilly, Speakers bureau: Orion Pharma, Biogen, Pfizer, CellTrion, Merck and Samsung Bioepis